instance store used case
Used case
A RAID mirror can be a good method to manage a disk level failure,
If the requirement is only for fast Read, the data does not change frequently,
If the data is changing, can consider file level or block level replication to a standby Instance with Instance Store volumes.
Possibly the design is no longer optimal and you should consider a High Performance Compute (HPC) solution which has additional safeguards and recovery methods built in.
as it provides time to mange the outage to move to a new instance.
However it alone is not a full solution as you will continue to run an a defective instance until it is fixed,
and it does not protect against other non-disk type failures.
Use S3 to host the master copy and basic snapshots or Lambda to pick up and write the incremental changes bask to S3.
AWS do not have a services for this at present, but it can be done with 3rd party products.
How to backup?
2 methods
You can back up data stored on an instance store volume in one of the following ways:
Create a new EBS volume, and then copy the data on your instance store volume to the EBS volume.
Back up the individual files stored on an EBS volum
such as rsync for Linux or robocopy for Windows.
rsync is CLI in Linux
Can form up-to-date / real-time snapshot to use to build a new instance.
swap space is fast
diagram
type of Ec2 supported
limitation
You can specify the instance store volumes for your instance only when you launch it. You can't attach instance store volumes to an instance after you've launched it.